Efficient Fabrication of TPU/MXene/Tungsten Disulfide Fibers with Ultra-Fast Response for Human Respiratory Pattern Recognition and Disease Diagnosis via Deep Learning.

Abstract

Flexible wearable pressure sensors have received increasing attention as the potential application of flexible wearable devices in human health monitoring and artificial intelligence. However, the complex and expensive process of the conductive filler has limited its practical production and application on a large scale to a certain extent. This study presents a kind of piezoresistive sensor by sinking nonwoven fabrics (NWFs) into tungsten disulfide (WS2) and Ti3C2Tx MXene solutions. With the advantages of a simple production process and practicality, it is conducive to the realization of large-scale production. The assembled flexible pressure sensor exhibits high sensitivity (45.81 kPa-1), wide detection range (0-410 kPa), fast response/recovery time (18/36 ms), and excellent stability and long-term durability (up to 5000 test cycles). Because of the high elastic modulus of MXene and the synergistic effect between WS2 and MXene, the detection range and sensitivity of the piezoresistive pressure sensor are greatly improved, realizing the stable detection of human motion status in all directions. Meanwhile, its high sensitivity at low pressure allows the sensor to accurately detect weak signals such as weak airflow and wrist pulses. In addition, combining the sensor with deep-learning makes it easy to recognize human respiratory patterns with high accuracy, demonstrating its potential impact in the fields of ergonomics and low-cost flexible electronics.